C Compilation Model

Comp-206 : Introduction to Software SystemsLecture 9

Alexandre DenaultComputer ScienceMcGill University

Fall 2006

Midterm

■ Date: Thursday, October 19th, 2006■ Time: from 16h00 to 17h30■ Content: Everything we have seen in class up to C

pointers. Unix operating system Shell Scripting Python C (including pointers)

■ Exact content of the midterm will be discussed in a latter class and posted on the web.

Servers - Sparcs

■ skinner.cs.mcgill.ca Hardware/OS: SunOS 5.8 sun4u sparc SUNW,Ultra-4 CPUs: 4 x 400 MHz (sparcv9 processors)

■ willy.cs.mcgill.ca Hardware/OS: SunOS 5.8 sun4u sparc SUNW,Ultra-80 CPUs: 4 x 450 MHz (sparcv9 processors)

■ nova.cs.mcgill.ca Hardware/OS: SunOS 5.8 sun4u sparc SUNW,Ultra-60 CPUs: 2 x 450 MHz (sparcv9 processors)

■ mimi.cs.mcgill.ca Hardware/OS: SunOS 5.8 sun4u sparc SUNW,Ultra-250 CPUs: 2 x 400 MHz (sparcv9 processors)

Servers - Intel

■ troy.cs.mcgill.ca Hardware/OS: Gentoo GNU/Linux running on a 2.6 kernel CPUs: 2 x 3.40 GHz (Intel Pentium 4 processors)

■ freebsd.cs.mcgill.ca Hardware/OS platform: FreeBSD 5.5-RELEASE-p3 CPUs: 2 x 3.40 GHz (Intel Pentium 4 processors)

Quiz

■ Give a regular expression that will match on the following: string “Quiz” line starting with string “Quiz”' or a digit line ending with string “Quiz” the string quiz, where the characters can be any case, e.g.,

QuIz, quiz, Quiz, etc. the string quiz, where it can be mis-spelled with K for Q and W

for U, e.g., kwiz, qwiz, etc. a string of at least 3 digits, starting with 7 lines containing no non-numeric characters, but at least one

numeric character.■ You have a directory containing a lot of files and

subdirectories, and you want to copy all of them except for the directory called big_dir. How do you do it?

History of C

■ The C programming language was created as a successor for B and BCPL.

■ It’s creation was parallel to the development of early Unix operating systems (1969-1973).

■ At the time, one of C’s strength was it’s portability.■ The first widely available description of the language

appeared in 1978,The C Programming Language (also known as the white book).

■ One of C’s most popular successor is C++, release in 1986.

Hello World

■ Traditionally, Hello World is the first application you write when starting with a new programming language.

#include <stdio.h>

int main(int argc, char *argv[]) { printf("Hello World"); return 0;}

Programming in C

From the users perspective, building a C program can be broken down in three steps:

■ Writing the source : Using an editor to write the source. You can use any text editor to write C code. Old-school C programmer often use Unix text editors such as

Vi or Vim. For large scale projects, an IDE (integrated development

environment) is preferable. Whatever editor you use, it should feature syntax highlighting

■ C programs are usually composed of several source files (we will take a look at this latter).

Programming in C (cont.)

■ The next step is to compile the program to a format the operating system can run.

■ A compiler is a program that translate a language to another. – A C compiler translates C code to machine code. – A Java compiler translates Java code to byte code.

■ For this course, we will use the GNU cc compiler (also known as gcc).

■ This compiler is installed on all the lab machines and servers.

Programming in C (cont.)

■ By default, the gcc compiler produces an executable files named a.out.

■ You can execute your program by running the a.out file. Don’t forget that a.out must be chmod executable. The

compiler usually takes care of this.■ Executable are compiled for specific architecture. If you

compile a program in the labs (Intel), it will not run on Mimi (Sun).

C Compilation Processor

Preprocessor

Compiler

Linker

Source code

Executable code

LibrariesAssembler

Preprocessor

■ The preprocessor is the first step of the compilation process.

■ It prepares the source files for the compiler.■ The preprocessor is responsible for . . .

Removing all the comments from the source files. Executing the preprocessor directives (#define and #include).

C Compiler

■ As previously mentioned, the compiler translate source code from one language to another.

■ The gcc compiler translate C code to assembler.■ Lets take the Hello World example.

#include <stdio.h>

int main(int argc, char *argv[]) { printf("Hello World"); return 0;}

Intel Assembly

main:pushl %ebpmovl %esp, %ebpsubl $8, %espandl $-16, %espmovl $0, %eaxsubl %eax, %espsubl $12, %esppushl $.LC0call printfaddl $16, %espmovl $0, %eaxleaveret

Sparc Assembly

main:save %sp, -112, %spst %i0, [%fp+68]st %i1, [%fp+72]sethi %hi(.LLC0), %o1or %o1, %lo(.LLC0), %o0call printf, 0nopmov 0, %i0retrestore

Assembler

■ The assembler takes assembly code and transforms it into object code.

■ Although object code is mostly composed of machine code, it cannot be executed by the operating system. Object code does not have the necessary references to

external functions and libraries to properly operate.

Linker

■ A linker takes the various outputs of a compiler and combines them to create an application. Sources files are compiled separately by the compiler. Those sources might reference a function that exists

elsewhere. The compiler leaves empty references to those functions. The linker fills those references using the compiled output of all

the files and the libraries available on the system.■ Once all the empty references have been resolved, the

linker combines all the compiler output to create an executable.

Libraries

■ C itself is a relatively small programming language.■ Most of it’s functionalities is provided through function

libraries. C provides a library for read/write to files and the screen. C provides a library to handle complicated math functionalities. C provides a library to retrieve the current time from the OS.

■ • A programmer is free (and encouraged) to use these libraries.

■ • The linker takes care to resolve references to library calls.

Gcc

■ As previously mentioned, Gcc is the Gnu C Compiler.■ Gcc encapsulates all the different step of the compilation

process. Create main.i, the preprocessed version of main.cgcc -E main.c

Create main.s, the assembler code of main.cgcc -S main.c

Create main.o, the object code of main.cgcc -c main.c

Create a.out, the compiled executable of main.cgcc main.c

Gcc options

■ -o filename : allows you to specify the name of the output executable (instead of a.out).

■ - v : enable verbose mode (more output information).■ -w : suppresses warning messages (bad idea)■ -W : extra warning messages (good idea)■ -Wall : all warning messages (best idea)■ -O1 : Optimize code for size and speed.■ -O2 : Optimize even more.

C vs Java - Similarities

■ C and Java have very similar syntax. Variable / function declarations Variable types : char, int, long, float, double Conditional statements : If, For, While

■ The notion of visibility is similar Variables declared in functions only exists in functions

C vs Java - Difference

■ C programming is much more low level Pointers and memory allocation

■ C is not object oriented No classes, no static methods, no interfaces. Libraries are completely different (no LinkedList, etc). Structures allow to group data together

■ C doesn't have Strings or boolean Strings are replaced by character arrays. boolean simply doesn't exist.

■ C is a single pass compiler Need to declare functions Header files

■ C has a preprocessor

C Functions

■ A C function has the same syntax as a Java function.

type function_name (parameters){ local variables C Statements

}

■ Functions have a return type, just like Java.■ However, unlike Java, they are not part of a class.■ In C, all functions behave as they were static.

Variables

■ Two types of variables exists in C Primitives Pointers

■ C primitives are very similar to Java primitives Char (1 byte, -127 to 128) Unsigned char (1 byte, 0 to 255) Short (2 bytes, -32768 to 32767) Int (4 bytes, -291 to 291 – 1) Float (4 bytes, ... ) Double (8 bytes, ... )

■ An unsigned variable is a numerical variable without a negative bit (thus allowing for larger numbers).

■ Notice there are no booleans or strings!

Global Variables

■ Variables not declared in a function are reference to as global.

■ Global variables can be accessed by any function in the program.

■ Global variables are very similar to static variable, only one copy exist.

■ Global variable should be avoided Since any functions can access global variable, it's difficult to

control access to those variable (an complicate debugging). They are not considered clean.